Reading Your First Report

When you run gostress --formats html, you get a dark-themed, interactive report with charts, metrics, and diagnostic insights. This guide walks through every section so you know exactly what you're looking at.

1. Hero Section

The top of the report shows your test headline and configuration pills:

• Method & URL — what was tested
• Workers — concurrency level
• Ramp — if --ramp was used, the ramp duration
• Target vs Actual duration — how long you asked for vs how long it actually ran

2. KPI Cards

Four summary metrics at a glance:

• Total Requests — every HTTP request sent during the test
• Success Rate — percentage of requests that returned a status code in your --success-status range (default: 200-399)
• Error Rate — the inverse: failed HTTP responses + transport errors
• Throughput — requests per second achieved, with utilization % if --rps was set

These numbers count up from zero when you open the report — a small animation, but it makes it easy to see which metrics are large vs small at a glance.

3. Reliability (Donut Chart)

The SVG donut chart visualizes the breakdown of your traffic:

• ● Green — Successful requests (status in success range)
• ● Amber — HTTP failures (status outside success range)
• ● Red — Transport errors (timeout, connection refused, etc.)

The center shows your overall success rate percentage. Hover over the legend items to highlight each segment.

4. Latency Distribution

Horizontal bar chart showing your latency percentiles:

• p50 (Median) — Half of requests finished faster than this. Your "typical" experience.
• p95 — 95% of requests finished faster. The slowest 5%.
• p99 — 99% of requests finished faster. Your tail latency.
• Average — Arithmetic mean. Can be skewed by outliers.
• Min / Max — Best and worst observed latency.

5. Status Codes

Bar chart + table showing every HTTP status code observed:

• Each bar is proportional to its count
• Color-coded: green for success codes, amber/red for errors
• The table below shows exact counts per code

If you see unexpected codes (e.g., 429 Too Many Requests), that's a signal to adjust your --rps limit or reduce concurrency.

6. Transport Errors

Errors that happen before the server sends a response:

• deadline_exceeded — Request timed out (default 2s)
• connection refused — Server rejected the TCP connection
• connection reset — Server dropped the connection mid-stream
• EOF / broken pipe — Server closed connection before sending a response

Below the chart, a diagnostic card explains what each error type means and provides contextual advice based on your actual error rate.

7. Throughput

Data transfer metrics:

• Total Data — total bytes received from the server
• Avg Throughput — MB/s transfer rate
• Utilization — if --rps was set, shows what % of the limit you actually hit

A "Analysis" note below adapts its message based on whether you used rate limiting and how close you got to the target.

8. Configuration & Results Tables

Two detailed tables at the bottom:

• Configuration — every flag and its value, including headers, payload, and success status spec
• Results — all raw numbers: request counts, latency stats, error breakdowns, timing

These are useful for copy-pasting into reports or comparing across runs.

Exporting as PDF

Click the Export PDF button in the top-right corner of the report. This opens your browser's print dialog — select "Save as PDF" to generate a clean, print-optimized version with a white background and proper page breaks.